Amazon’s latest healthcare foray is a major step into health informatics
A new software service from the company called Amazon Comprehend Medical uses machine learning and natural language processing to extract relevant medical information from doctors’ notes, clinical trial reports and patient health records.
Amazon’s steady march into the healthcare industry has reached a new milestone with the development of HIPAA-eligible machine learning software that can accurately digitize and process medical records.
Dubbed Amazon Comprehend Medical, the service uses natural language processing to scan and extract relevant information from doctors’ notes, clinical trial reports and patient health records.
“There are no servers to provision or manage – developers only need to provide unstructured medical text to Comprehend Medical. The service will read the text and then identify and return the medical information contained within it,” the company wrote in a blog post announcing the technology.
While the shift to EHRs created a foundational data layer to potentially build technology on top of, there’s still been significant challenges in effectively utilizing that information.
A major problem stymieing analysis of medical records is that the information is often unstructured, messy and difficult to parse through.
Amazon claims that its technology, aided by machine learning, can “reliably understand the medical information in unstructured text, identify meaningful relationships, and improve over-time.”
One simple example is being able to understand that the condition known as “methicillin-resistant Staphylococcus aureus” and “MRSA” are equivalent and identify whether the patient tested positive for the infection.
Amazon says that the tool, which is available through an API call, can both highlight important medical information in records and identify relevant relationships. For example being able to understand how a diagnosis for a condition is related to a specific medication prescription and dosage.
A focus by Amazon is on data privacy for protected health information, which has often remained a barrier for large scale data analysis and transfer in the industry.
The company says the software can quickly identify PHI and can be used to process, store and transmit the information while also adhering to European GDPR standards.
Amazon Web Services says it doesn’t store or save healthcare data uploaded to its platform, which is encrypted and only available to customers themselves.
Potential use cases include more efficient and accurate medical coding, clinical trial management, the development of population health platforms and creating better data-enabled support for clinical decisions.
The new technology represents a big push by Amazon into the multi-billion dollar health informatics market already populated by big-name public companies like UnitedHealth Group’s Optum division and Inovalon as well as startups like San Mateo-based Apixio and San Francisco-based Talix.
This year has seen the e-commerce giant progressively ramp up its ambitions in healthcare with moves like its $1 billion PillPack acquisition, its health collaboration with J.P. Morgan and Berkshire Hathaway and the addition of big name healthcare players to its roster.
In fact, one of the executives leading the development of Amazon Comprehend Medical has been Dr. Taha Kass-Hout, the former chief health informatics officer for the FDA, who joined Amazon earlier this year after a stint at Trinity Health.
Early customers of Amazon Comprehend Medical include major healthcare stakeholders including pharma giant Roche, the Fred Hutchinson Cancer Research Center and consulting and financial services firms like Deloitte and PWC.
Fred Hutchinson is using the software for clinical trial management and recruitment, scanning millions of clinical notes for specific medical conditions, medications and treatments.
“The process of developing clinical trials and connecting them with the right patients requires research teams to sift through and label mountains of unstructured clinical record data,” said Fred Hutchinson CIO Matthew Trunnell in a statement.
“Amazon Comprehend Medical will reduce this time burden from hours to seconds. This is a vital step toward getting researchers rapid access to the information they need when they need it so they can find actionable insights to advance lifesaving therapies for patients.”
The software has a pay-as-you-go pricing model allows customers to upload their health records to AWS and get organized information in return at a cost of up to a penny per 100 characters analyzed, depending on the data desired.